Potentiation of the activity of the gamma-aminobutyric acid type A (GABA(A)) receptor channel by volatile anesthetic agents is usually studied in vitro at room temperature. Systematic variation of temperature can be used to assess the relevance of this receptor to general anesthesia and to characterize the modulation of its behavior by volatile agents at normal body temperature.
Potentiation of the GABA(A) receptor by halothane, sevoflurane, isoflurane, and methoxyflurane was studied at six temperatures in the range 10-37 degrees C using the whole-cell patch-clamp technique and mouse fibroblast cells stably transfected with defined GABA(A) receptor subunits.
Control GABA concentration-response plots showed small and physically reasonable changes in the GABA concentration required for a half-maximal effect, the Hill coefficient, and maximal response over the range 10-30 degrees C. Potentiations of GABA (1 microM) responses by aqueous minimum alveolar concentrations of the volatile anesthetic agents decreased with increasing temperature from 10-37 degrees C in an agent-specific manner (methoxyflurane > isoflurane > sevoflurane > halothane) but tended to equalize at normal body temperature (37 degrees C). These findings are in line with published results on the temperature dependence of anesthetic potencies in animals.
These results are consistent with direct binding of volatile anesthetic agents to the GABA(A) receptor channel playing an important role in general anesthesia. The finding that the degree of anesthetic potentiation was agent-specific at low temperatures but not at 37 degrees C emphasizes the importance of doing in vitro experiments at normal body temperature.
THE [small gamma, Greek]-aminobutyric acid type A (GABAA) receptor channel is widely thought to be a major target for most general anesthetic agents. [1,2]This neurotransmitter-gated receptor Cl-channel is not only distributed throughout the mammalian central nervous system, but its GABA-activated inhibitory actions are substantially potentiated by clinically relevant concentrations of a diverse range of inhalational and intravenously administered general anesthetic agents (ketamine appears to be an exception ). Nonetheless, although it is clear that certain intravenously administered anesthetic agents, such as etomidate [5,6]and propofol, act rather specifically on GABAAreceptors, it is often argued that nonspecific volatile agents, such as isoflurane and halothane, produce general anesthesia by acting at a large number of molecular targets. One way to approach this problem is to make use of the finding that mammalian values for minimum alveolar concentration (MAC) for volatile agents are dependent on body temperature, with the degree of temperature dependency varying from agent to agent. [8-11]If the GABAAreceptor is a major target for volatile anesthetic agents, then one might expect temperature effects on MAC to be mirrored by temperature effects on the potentiation of the GABAAreceptor by these agents. Further, at normal body temperature, the potentiations by MAC values of different volatile agents should be comparable. In this article, we describe experiments that test these predictions, using electrophysiologic techniques at controlled temperatures on cells stably transfected with defined GABAAsubunits.
The results of these temperature experiments also are pertinent to the question of whether volatile anesthetic agents act on the GABAAreceptor by binding directly to the receptor protein or indirectly by dissolving in the surrounding lipid bilayer, a problem that has been approached elsewhere using optical isomers of the volatile agent isoflurane [12-14]and the techniques of molecular genetics. [15,16]Finally, the effects of volatile anesthetic agents on the GABAAreceptor channel observed at reduced temperatures may be of interest to those who use these agents to perform operations during hypothermic conditions (e.g., cardiac surgery during cardiopulmonary bypass).
Materials and Methods
Cells were prepared and cultured as described previously. Mouse fibroblast L-cells stably transfected with complementary deoxyribonucleic acids (cDNAs) encoding geneticin resistance, and GABAAbovine [small alpha, Greek]1, [small beta, Greek]1, and [small gamma, Greek]2Lreceptor subunits were kindly supplied by P. Whiting (Merck Sharp and Dohme Research Laboratories, Harlow, Essex, UK) at passage number 10. In these PA3 cells, expression of GABAAreceptors is under the control of a hormone-sensitive promoter; without addition of the hormone dexamethasone to the culture medium, there is little or no expression. PA3 cells were cultured in Dulbecco's modified Eagle's medium supplemented with heat-inactivated fetal bovine serum (10% vol/vol), L-glutamine (2 mM), penicillin (100 U/ml), and streptomycin (100 mg/ml)(all obtained from Gibco Life Technologies, Paisley, UK). Unless otherwise stated, all other chemicals were obtained from Sigma Chemical Co. (Poole, Dorset, UK). Cells were maintained in an atmosphere of 95% air/5% CO2and 95% relative humidity at 37 [degree sign]C. The PA3 cells were passaged every week using (nonenzymatic) Cell Dissociation Solution (Sigma Chemical Co.) and seeded to give a passage ratio of [almost equal to] 1:10. As the passaged cells approached confluence (usually after 4 days), they were exposed to culture medium containing the antibiotic agent geneticin (2 mg/ml) to select for cells with geneticin resistance. For the whole-cell patch-clamp electrophysiology experiments, PA3 cells were grown on (uncoated) glass coverslips (6 x 8 mm) in six-well dishes. The culture medium was supplemented with 1 [micro sign]M dexamethasone, and cells were incubated for >or= to 2 days before conducting electrophysiologic experiments. PA3 cells with passage numbers ranging from 12 to 34 were used for all experiments, with no noticeable differences in the levels of expression.
Inward currents (because of the efflux of Cl-ions) evoked by the application of GABA to PA3 cells and patch membranes were recorded using the whole-cell and outside-out patch-clamp techniques with an Axopatch 200 amplifier (Axon Instruments, Foster City, CA). Recording pipettes were fabricated from thin- and thick-walled filamented borosilicate glass capillary tubes (GC150TF and GC150F, respectively; Clark Electromedical Instruments, Reading, Berkshire, UK) using a two-stage pull (Narishige PB-7 micropipette puller; Narishige, Tokyo, Japan). Pipettes were filled with intracellular recording solution containing 130 mM CsCl, 1 mM MgCl2, 10 mM HEPES, and 11 mM EGTA (titrated to pH 7.2 with CsOH). During recording, the PA3 cells were bathed in an extracellular solution containing 124 mM NaCl, 5 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 5 mM HEPES, and 11 mM D-glucose (titrated to pH 7.4 with NaOH). Currents were low-pass filtered (8-pole Bessel filter; Frequency Devices, Inc., Haverhill, MA) and digitized (Axon Instruments TL-1 interface) using pCLAMP5 software and stored on a hard disk for off-line analysis.
For whole-cell recording from PA3 cells, thin-walled recording pipettes, lightly fire-polished to give typical electrode resistances of 2-5 M Omega, formed giga-ohm seals with the cells. Once the whole-cell configuration was established, the cells were voltage-clamped at -40 mV. Membrane currents were filtered at 50 Hz (-3 dB) and sampled at 100 Hz. Series resistance was compensated for by >80% in all recordings.
For single-channel recording from PA3 cell membranes, outside-out patches were pulled from PA3 cells grown on poly-L-lysine-coated coverslips. Thick-walled recording pipettes were fire-polished to give typical electrode resistances of [tilde operator] 25 M Omega. Once the outside-out patch configuration was established, the patch was voltage clamped at -120 mV. Membrane currents were filtered at 2 kHz (-3 dB) and sampled at 4 kHz.
All drugs were applied rapidly to the cells via a double-barrelled tube (two fused glass capillaries, each with an inside diameter of 1 mm) positioned [almost equal to] 400 [micro sign]m upstream from the cell as described previously. 
The temperature of the cells was controlled by delivery of cooled or warmed solutions via the double-barrelled capillary tubes and the bath. The polytetrafluoroethylene tubing of the perfusion system was passed (195 mm total path length for each solution) through a copper block connected to a heatsink via two 33.4-W Peltier effect pumps (R. S., Corby, Northants., UK). The temperature <500 [micro sign]m downstream from the cell was recorded using a thermocouple probe connected to a digital thermometer (Digitron Instruments 1408K, Hertford, Herts., UK). Using this apparatus, the temperature of the cells could be set between 10 and 40 [degree sign]C with a precision of 0.2 [degree sign]C and a maximum equilibration time of <2 min.
Volatile general anesthetic agents were obtained from Abbott Laboratories Ltd., Queenborough, Kent, UK (methoxyflurane, isoflurane, and sevoflurane) and Aldrich Chemical Co. Ltd., Gillingham, Dorset, UK (halothane). Anesthetic solutions were prepared by diluting saturated aqueous solutions made up at room temperature (20-23 [degree sign]C); the concentrations of these saturated solutions were taken to be 9.1 mM methoxyflurane, 15.3 mM isoflurane, 11.8 mM sevoflurane, and 17.5 mM halothane. During electrophysiologic experiments, cells were exposed to 1 mammalian MAC (at 37 [degree sign]C) of anesthetic agent: 340 [micro sign]M methoxyflurane, 280 [micro sign]M isoflurane, 330 [micro sign]M sevoflurane, and 230 [micro sign]M halothane. These values were determined by taking the means of tabulated human, dog, and rat (or mouse) aqueous MAC values. [24,25]Reservoirs containing volatile anesthetic agents were sealed with rigid plastic floats, and all tubing and valves were made of polytetrafluoroethylene.
The effects of the four volatile general anesthetic agents were recorded on currents activated by 1 [micro sign]M GABA; this low concentration activated nondesensitizing currents that were [almost equal to] 1% of a maximum response. In each determination, pairs of control currents were recorded before and after two test exposures to anesthetic agent. The mean of these four control currents defined the control response. To elicit the test currents between these controls, cells were exposed to anesthetic agent and GABA together; there was no preexposure to the anesthetic agent (the anesthetic agent alone had no effect on the baseline current at the low concentrations used here). The mean of the two test currents in the presence of anesthetic agent defined the test response.
Analysis of Whole-cell Currents
GABA concentration-response data at each temperature were fitted (unweighted least-squares) to a Hill equation:Equation 1where E is the peak of each GABA-induced current expressed as a percentage of the maximal peak current, A is the GABA concentration, nHis the Hill coefficient, and EC50is the GABA concentration required for a half-maximal effect.
To measure the effect of temperature on the maximal control current, the following approach was used to avoid problems associated with recording currents activated by saturating concentrations of GABA (such as desensitization). Cells were exposed to 10 [micro sign]M GABA at 10, 20, and 30 [degree sign]C and the currents measured (nine observations at each temperature from at least eight cells). The mean of pairs of currents activated by 10 [micro sign]M GABA at 20 [degree sign]C before and after a temperature change defined the reference response; between these reference currents, cells were cooled to 10 [degree sign]C or warmed to 30 [degree sign]C and the size of the currents relative to the reference response recorded. Because the GABA EC50and Hill coefficient values had been determined at all three temperatures, the relative currents activated by 10 [micro sign]M GABA could be used to calculate the maximum response.
Percent anesthetic potentiation was defined as 100%(I - Io)/I (o), where Iois the mean control GABA response and I is the mean test GABA response (i.e., in the presence of anesthetic agent).
Analysis of Single-channel Currents
Current records were divided into 2,048 sample episodes. An episode containing single-channel events was used to construct a current amplitude all-points histogram (bin width, 0.05 pA). Two gaussians were fitted to the current histogram: one to the data representing the open state and one to the data representing the closed state of the channel (least-squares method). The mean open channel current was determined from the difference between the means of the two gaussians.
All values are given as the mean +/− SEM.
GABA-induced whole-cell currents in PA3 cells were first studied during control conditions (in the absence of anesthetic agent) at three different temperatures (10, 20, and 30 [degree sign]C) and six different GABA concentrations (1, 3, 10, 30, 100, and 300 [micro sign]M). As expected for GABAAreceptor channel Cl-currents and the concentrations of chloride used in our intracellular and extracellular solutions, these GABA-induced currents reversed close to 0 mV (range, 0.4-2.8 mV) at all temperatures. At the standard holding potential of -40 mV, currents were always inward, corresponding to a net efflux of negatively charged Cl-ions from the cells.
GABA concentration-peak response data at each temperature fit well to a Hill equation (Equation 1). When expressed as a percentage of the maximum current at each temperature, the three Hill curves differed minimally from each other (Figure 1). When examined more closely, it can be seen that, with rising temperature, there was a slight increase in the GABA EC50value (Figure 2A) and a small decrease in nH(Figure 2B). These changes were linear with temperature and are consistent with small decreases in the apparent affinity and cooperativity of GABA binding to the GABAAreceptor channel with increasing temperature. There was also a linear increase with increasing temperature of the maximal GABA-induced whole-cell current (Figure 2C) and of the single-channel conductance from outside-out patches (Figure 2D). These increases are similar to each other and to that for diffusion of Cl-ions in water (see Discussion), suggesting that they are attributable to increasing temperature increasing Cl-mobility within the open aqueous pore of the GABAAreceptor channel.
Thus, the effects of temperature on the activity of the GABAAreceptor channel in the absence of general anesthetic agents were small, linear, and physically reasonable, encouraging us to look for temperature-dependent effects of anesthetic agents on this system. It is well known that a wide variety of general anesthetic agents potentiates the GABA-induced activity of the GABAAreceptor, because of an apparent increase in GABA affinity. We chose to study the temperature dependence of potentiation at a low concentration of GABA (1 [micro sign]M) by aqueous MAC values (at 37 [degree sign]C) of four volatile agents (halothane, isoflurane, methoxyflurane, and sevoflurane). A low GABA concentration was used to maximize the degree of potentiation without causing desensitization, and aqueous MAC values at 37 [degree sign]C were used because of their pharmacologic relevance.
Typical current traces at four temperatures (10, 20, 30, and 37 [degree sign]C) in the presence and absence of the four volatile agents are shown in Figure 3. In all cases, the anesthetic agent potentiated the response to GABA. As expected, the activation and deactivation kinetics were faster at the higher temperatures, in the presence and absence of anesthetic agent. More importantly, the degree of anesthetic potentiation varied from agent to agent but decreased with increasing temperature. This latter feature can be seen more clearly in Figure 4, where the mean percent potentiations for all four anesthetic agents at six different temperatures are plotted. It is noteworthy that anesthetic potentiations were remarkably different and agent-specific at low temperatures (e.g., methoxyflurane much > isoflurane > sevoflurane >or= to halothane at 10 [degree sign]C), but they decreased and tended to converge (to [almost equal to] 100 - 200% potentiation) as the temperature was increased to 37 [degree sign]C (the typical mammalian body temperature).
To our knowledge, this is the first electrophysiologic study of the effects of volatile anesthetic agents and temperature on the GABAAreceptor channel. This is surprising in view of the probable importance of this neurotransmitter-gated ion channel in general anesthesia and is perhaps attributable to the experimental difficulties of working with volatile agents over a wide range (10-37 [degree sign]C) of temperatures.
Even during control conditions, in the absence of an anesthetic agent, there have been few electrophysiologic investigations into how temperature affects the mammalian GABAAreceptor channel. We measured GABA concentration-dependent responses as a function of temperature. We found (Figure 1and Figure 2) that with increasing temperature from 10-30 [degree sign]C there was an increase in the GABA EC50value and the maximal peak current response but a decrease in nH. These effects were small and, to the accuracy of our data, linear, which is consistent with no major temperature-dependent changes in the structure or functioning of the receptor channel. In addition, the observed effects were physically reasonable. The effects of temperature on the GABA EC50(Figure 2A) and nH(Figure 2B) values are consistent with GABA binding less tightly and with reduced cooperativity with increasing temperature, whereas the increase with temperature of the maximal peak current (Figure 2C) is what might be expected for the temperature dependence for Cl-ions diffusing through the aqueous pore of the channel. Increasing temperature was found to increase single-channel conductance (Figure 2D) in a manner quantitatively similar to the effects of temperature on maximal peak current and the diffusion coefficient for Cl-in bulk aqueous solution. This result (Figure 2C and Figure 2D) suggests that the mobility of Cl (-) anions within the aqueous pore of the GABAAreceptor channel may be similar to that in water, a property that might be necessary to ensure rapid fluxing of Cl-.
In contrast to the small effects of temperature on control GABA responses, the temperature dependence of anesthetic modulation was large for some volatile agents (Figure 3and Figure 4). For example, potentiation of GABA currents by a fixed aqueous concentration of methoxyflurane (340 [micro sign]M = MAC) increased from 200% at 37 [degree sign]C to 730% at 10 [degree sign]C, whereas potentiation by isoflurane (280 [micro sign]M = MAC) increased from 90% at 37 [degree sign]C to 380% at 10 [degree sign]C. In contrast, potentiation by aqueous MAC values of halothane (230 [micro sign]M) and sevoflurane (330 [micro sign]M) increased much less over the same temperature range (Figure 3and Figure 4). This division of volatile agents into two groups (methoxyflurane and isoflurane vs. halothane and sevoflurane) on the basis of their temperature effects on GABAAreceptor potentiation is reminiscent of their grouping (methoxyflurane and isoflurane vs. halothane) according to the temperature dependence of their aqueous MAC values for producing general anesthesia in dogs and mice. (We are not aware of any studies on how the MAC of sevoflurane changes with temperature.) In both cases (GABAAreceptors and whole animals), changes in temperature affect the behavior of methoxyflurane and isoflurane considerably more than that of halothane, with all three agents being more effective at low temperatures.
In addition, it is striking that the large agent-specific differences in anesthetic potentiation of the GABAAreceptor channel observed at low temperatures progressively decrease with increasing temperature and almost disappear at the normal body temperature of 37 [degree sign]C (Figure 4). (The slightly larger potentiation at 37 [degree sign]C by methoxyflurane may reflect the large animal variations in estimates of MAC for this agent.) Not only is this what one would expect of a receptor that plays a major role in the general anesthesia produced by volatile agents, but it also illustrates the importance of working at physiologic body temperatures. For example, had we performed experiments only at room temperature (20 [degree sign]C), we might have predicted erroneously (Figure 4) that isoflurane is considerably more effective than halothane at potentiating GABA responses in the normal animal.
Our temperature results also bear on the question of how volatile anesthetic agents modulate GABAAreceptor activity. Do these agents bind directly to the receptor (the protein hypothesis [28,29]), or do they influence its activity secondarily to binding to and perturbing the surrounding lipid bilayer (the lipid hypothesis )? We found that decreasing temperature increases the effectiveness of a fixed aqueous concentration of volatile agent at the GABAAreceptor (a similar result has been obtained with a neuronal nicotinic acetylcholine receptor ), whereas partitioning of isoflurane and halothane from water into lipid bilayers decreases with decreasing temperature. This means that, with decreasing temperature, potentiation of GABAAactivity by these agents increases even though less anesthetic agent is actually in the lipid bilayer. This is difficult to reconcile with the lipid hypothesis but is consistent with the protein hypothesis. The temperature dependence observations for volatile anesthetic potentiation of the GABAAreceptor channel are consistent with these agents binding directly to the receptor channel protein and mirror the temperature dependence of their potencies for mammalian general anesthesia.
The authors thank the Medical Research Council for its support and Dr. P. Whiting for the gift of the stably transfected PA3 cells.